This prospective cohort study quantified the longitudinal associations between long-term exposure to multiple air pollutants and cognitive impairment risk among China's middle-aged and older adults. Leveraging data from 10,584 participants in the China Health and Retirement Longitudinal Study (2011-2020), we assessed high-resolution (1-km) annual exposures to particulate matter (PM, PM, PM), nitrogen dioxide (NO), and ozone (O) using satellite-based models. Cognitive impairment was defined through education-stratified Chinese Mini-Mental State Examination thresholds. Time-varying Cox proportional hazards models revealed significant dose-response relationships: per 10-μg/m³ increase, PM(HR=1.078, 95 % CI:1.022-1.136), PM(HR=1.042, 1.015-1.070), PM(HR=1.038, 1.022-1.054), and NO(HR=1.129, 1.061-1.200) elevated cognitive impairment risk, while Oshowed no association (HR=0.878, 0.743-1.038). Three-year exposure windows intensified effects for NO(HR=1.053) and PM(HR=1.034). Multi-pollutant models demonstrated PM's exceptional robustness (e.g., PM+NOHR=1.055), whereas NOattenuated finer particulate associations. Critical vulnerability emerged among adults with education below junior high (PMHR=1.30), rural residents (PMHR=1.21), those aged 60-69 years (peak PMHR=1.28), and uninsured individuals (PMHR=1.22). Strictly linear exposure-response relationships (all Pnonlinear>0.05) indicated no safe thresholds. These findings advocate prioritizing control of PM(particles ≤10 μm) alongside mitigation of traffic-related pollutants, while implementing targeted interventions for vulnerable educationally disadvantaged and rural populations to mitigate dementia burden. 1 2.5 10 2 3 1 2.5 10 2 3 2 1 10 10 2 2 1 10 1 1 10
